Why Your Laser Stops Cutting Mid-Job (It's Not the Laser)
Your fiber laser stops cutting halfway through a big job, or leaves missing passes. Before you blame the machine, check these common, often overlooked, settings and causes from our jewelry workshop.
If your fiber laser suddenly stops cutting halfway through a big job, or seems to be leaving missing passes and undercutting, it is almost certainly not a problem with the laser source itself. In our workshop, we have lost countless pieces of silver to these exact symptoms, only to find the root cause was a subtle setting or a hidden software issue.
Here, we will walk you through the most likely culprits, from the most common to the easiest to overlook, so you can diagnose why your machine is failing to cut through.
The Controller Buffer: Your Passes Are Being Silently Dropped
This is, by far, the most insidious problem when a big job fails to cut through, and it looks exactly like your laser suddenly got weaker. Your PC sends a stream of commands to the JCZ controller board, but that board has a finite buffer. On our board, it is a couple of megabytes, enough for a few hundred command lists.
When you are running a very large cut – many passes over complex geometry – it is possible for your computer to send commands faster than the board can process them and mark the metal. Once the board's buffer is full, it cannot accept new commands and will silently drop them.
What does this mean for your job? You might have programmed 39 passes, but the metal only received 25. The machine seems to go quiet long before the software says the job is done, and adding passes helps less than it should, because many of those new passes are also getting dropped.
We saw this plainly in our workshop: a cut that should have taken 147 seconds of marking time only fired for about 25 seconds. The operator's natural reaction is to just add more passes, which hides the underlying bug and wastes both time and expensive material.
The solution is proper flow control: the software driving the laser must never send more than a few seconds of marking ahead of what the board has actually played. Once we implemented this in our own software, the number of passes we needed for the same cuts dropped dramatically, because all the commanded passes were finally reaching the board.
Check for these symptoms if your laser stops mid job or shows missing passes:
- Big jobs cut worse per pass than smaller ones.
- The machine goes quiet and finishes marking much faster than the software timer indicates.
- Adding more passes helps, but not as much as you would expect, or it feels like you need an unreasonably high number of passes.
Focus Height: A Millimeter Can Kill Your Cut
After the controller buffer, incorrect focus height is the next biggest culprit for cutting failures. Even a small error can completely prevent the laser from cutting through.
A fiber marking laser is a Class 4 device. The invisible 1064 nm beam and its reflections cause instant, permanent eye damage. Laser safety glasses rated for 1064 nm (with an appropriate Optical Density for your source power) must be worn by everyone in the room, always. Never leave a running job unattended. Metal marking produces fine particulate; proper extraction and filtration are essential health equipment.
The factory-calibrated focus height for your machine is stored in EzCad3's Motors.ini file as m_dFocusPos. On our machine, this value is -59.6 mm, which is the Z-axis position where the surface of the machine's platform is in perfect focus.
When you cut a sheet of material on the platform, the top surface of that material is one thickness above the platform. Therefore, the rule for setting your Z-axis height is: Z = focus position + material thickness.
For example, to cut 1 mm thick sterling silver on our machine, the correct Z-axis height would be -59.6 mm + 1 mm = -58.6 mm.
We learned this the hard way: a Z-axis setting of -58.5 mm cuts, but -56.5 mm (just 2 mm higher) does not cut at all. A small error of ±0.5 mm might be barely noticeable, but 2 mm of defocus completely kills the cutting power. For months, our own software needed far more passes than EzCad3 for the same job, simply because our Z-axis reference was about 2.1 mm too high.
If your cuts have gotten worse over time and nothing in your design file or parameters has changed, check your focus height before you adjust power or pass counts. A bent jig, a thicker batch of material, or a platform that has shifted slightly can all lead to "the laser got weaker" symptoms.
Wobble Settings: The Invisible Wall
Cutting metal with a fiber laser relies heavily on a setting called "wobble." This causes the laser beam to oscillate slightly sideways as it follows the cut line. This creates a kerf (the width of the cut) that is wider than the beam's spot size, allowing molten metal to escape instead of re-welding behind the beam.
On our machine, for cutting sterling silver, our working recipe uses a wobble diameter of 0.15 mm, an end diameter of 0.05 mm, and a distance of 0.07 mm.
Here is where a subtle error can ruin everything: a wobble setting that is too small simply will not cut at all. We once had the wobble diameter set to 0.015 mm instead of 0.15 mm – a factor of ten error, a single misplaced decimal point. A 15 micron wobble is effectively a thin line. The machine would mark the silver perfectly, but the cut line would never open up, no matter how many passes we ran. Everything else in the recipe – power, speed, frequency, passes – was correct, and the machine simply would not go through.
If a recipe that used to cut reliably suddenly only engraves a line and fails to cut through, check your wobble diameter setting immediately. This is one of the most misleading failures because the job looks like it is running exactly as it should, just without cutting.
Not Enough Passes (or the Wrong Frequency)
Sometimes, the problem is simply that you are not commanding enough passes to cut through. This is especially true with sterling silver.
The Challenge of Sterling Silver
A fiber laser at 1064 nm must be absorbed by the material to do its work. Sterling silver is one of the most reflective metals there is, and an excellent heat conductor. This means it rejects much of the laser beam and quickly carries away any heat that does get absorbed. That is why it typically requires many passes to cut through.
For more detail on our measured silver cutting data, you can read How Many Passes to Cut Silver? Our Measured Curve.
Our workshop's common recipe for cutting sterling silver sheet on the platform is 95 % power, 190 mm/s, and 60 kHz, with wobble enabled. With these settings, our measured pass counts are:
- 0.8 mm silver: 17 passes
- 0.9 mm silver: 21 passes
- 1.0 mm silver: 28-29 passes
These numbers are specific to our machine, our source, and our silver. Always treat them as a starting point and prove them on scrap material.
Pass count, not raw power, is usually the best lever for getting through thick material. Turning the power up on a job that will not cut often just widens the kerf and heats the piece, leading to distortion.
Frequency for Cutting vs. Engraving
Frequency behaves differently for cutting and engraving. For deep engraving, lower frequencies (fat, powerful pulses) are often preferred. However, for cutting through, our tests on 0.3 mm silver showed the opposite: dropping the frequency to 20 kHz made the cut worse, requiring 16 passes and still not going through. Back at 60 kHz, it cut in a handful of passes. High frequency (more, smaller pulses) wins for cutting through; low frequency wins for surface ablation depth.
Cutting Gold: A Different Experience
If you have only ever cut silver, cutting gold will feel like your machine suddenly got much better. Gold reflects less of the 1064 nm beam than silver, making it significantly easier to cut and engrave with the same machine and settings.
The base recipe (95 % power, 190 mm/s, 60 kHz, wobble on) does not change for gold, only the number of passes. Our gold table works out to roughly half the passes we use for silver of the same thickness. These are derived numbers we use as a starting point, not directly measured data, and must be verified on scrap.
Always start with fewer passes and work your way up. A cut that does not go through costs you one more pass; a cut with too many passes will widen the kerf, overheat the piece, and distort thin sections. Gold is expensive to learn on.
This table assumes a typical yellow gold. Karat and alloy dramatically change how a material behaves. 18k yellow is not 14k, and white gold, due to its alloying metals, will behave differently again.
| Sheet thickness | Gold passes (starting point, derived) |
|---|---|
| 0.25 - 0.4 mm | 2 |
| 0.5 mm | 3 |
| 0.6 mm | 4 |
| 0.7 mm | 6 |
| 0.8 mm | 9 |
| 0.9 mm | 11 |
| 1.0 mm | 14 |
Practical Takeaways: What to Check First
When your laser suddenly stops cutting or gives you inconsistent results on a big job, here is the order in which we would troubleshoot:
- Check for Controller Buffer Issues: If you are seeing missing passes, or the job is finishing much faster than expected, this is your prime suspect. Look for the symptoms described above, particularly on larger jobs.
- Verify Focus Height: This is the most common mechanical error. A small change in material thickness, a bent jig, or a shifted platform can throw your focus off by enough to prevent cutting. Re-measure your Z-axis height for your material's specific thickness.
- Confirm Wobble Settings: If the job is marking perfectly but never cutting through, double-check your wobble diameter. A misplaced decimal point can make the wobble effectively disappear, trapping the melt and preventing a through-cut.
- Increase Pass Count (with caution): If all the above are correct, you might simply need more passes for your material and desired depth. Remember to use the correct high frequency for cutting. Always test on scrap before committing to a valuable piece.
By systematically checking these points, you should be able to diagnose and fix most instances of a fiber laser that stops cutting mid-job, saving you material, time, and frustration.